Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
| Author |
Topic |
|
adrianmuller
Starting Member
4 Posts |
Posted - 2009-04-07 : 07:50:02
|
| HelloI need to know the average number of days between the first and second click, the second and third, third and fourth click and so on for different users that each have a row every time they click.Excerpt from data:Computer ID Click Date111 03/10/2008111 08/10/2008111 18/10/2008111 03/11/2008111 03/12/2008222 03/10/2008222 08/10/2008So for example user 111, I need to know difference between 1st & 2nd, 2nd & 3rd, 3rd & 4th and 4ht & 5ht, while for user 333 I only need the difference between 1st & 2nd.Problems that I have:-I can not use only the max and min functions, cause there are in between values I also need to estimate. -Users have a different number of clicks, so they are in one or more rows.I want to group by computer id, and be able to tell, the average number of days between 1st and 2nd click for all users is....? the average number of days between 2nd and 3rd click for all users is...?Thanks in advanced |
|
|
SwePeso
Patron Saint of Lost Yaks
30421 Posts |
Posted - 2009-04-07 : 08:06:37
|
[code]DECLARE @Sample TABLE ( ComputerID INT, ClickDate DATETIME )SET DATEFORMAT DMYINSERT @SampleSELECT 111, '03/10/2008' UNION ALLSELECT 111, '08/10/2008' UNION ALLSELECT 111, '18/10/2008' UNION ALLSELECT 111, '03/11/2008' UNION ALLSELECT 111, '03/12/2008' UNION ALLSELECT 222, '03/10/2008' UNION ALLSELECT 222, '08/10/2008';WITH Yak (ComputerID, ClickDate, recID)AS ( SELECT ComputerID, ClickDate, ROW_NUMBER() OVER (PARTITION BY ComputerID ORDER BY ClickDate) FROM @Sample)SELECT y1.ComputerID, y1.ClickDate, DATEDIFF(DAY, y2.ClickDate, y1.ClickDate) AS [Days]FROM Yak AS y1LEFT JOIN Yak AS y2 ON y2.ComputerID = y1.ComputerID AND y2.recID + 1 = y1.recID[/code] E 12°55'05.63"N 56°04'39.26" |
 |
|
|
ddramireddy
Yak Posting Veteran
81 Posts |
Posted - 2009-04-07 : 08:13:21
|
| Assume that your table is like thiscreate table ComputerClicks( [Computer ID] int , [Click Date] datetime, [UserId] int)then the below query will give your required output.with cte as ( select *,row_number() over (partition by [Computer Id],UserId order by [Click Date]) as rn from ComputerClicks ),cte1 as (select c.[Computer Id],datediff(day,c.[Click Date],c1.[Click Date]) as Days,'Between ' + cast( c.RN as varchar) + ' and ' + cast( c1.RN as varchar) as CLickType from cte cinner join cte c1 on c1.rn = c.rn + 1 and c.[Computer Id] = c1.[computer Id] and c.UserId = c1.UserId)select [Computer Id],ClickType,Avg(Days) from cte1group by [Computer Id],ClickType |
 |
|
|
adrianmuller
Starting Member
4 Posts |
Posted - 2009-04-07 : 09:44:03
|
| I am going to try it out and will let you know. thanks to both |
 |
|
|
adrianmuller
Starting Member
4 Posts |
Posted - 2009-04-07 : 10:18:29
|
Dear ddramireddy,Why do I have to create a join, the info is already in just one table. quote: Originally posted by ddramireddy Assume that your table is like thiscreate table ComputerClicks( [Computer ID] int , [Click Date] datetime, [UserId] int)then the below query will give your required output.with cte as ( select *,row_number() over (partition by [Computer Id],UserId order by [Click Date]) as rn from ComputerClicks ),cte1 as (select c.[Computer Id],datediff(day,c.[Click Date],c1.[Click Date]) as Days,'Between ' + cast( c.RN as varchar) + ' and ' + cast( c1.RN as varchar) as CLickType from cte cinner join cte c1 on c1.rn = c.rn + 1 and c.[Computer Id] = c1.[computer Id] and c.UserId = c1.UserId)select [Computer Id],ClickType,Avg(Days) from cte1group by [Computer Id],ClickType
|
 |
|
|
visakh16
Very Important crosS Applying yaK Herder
52326 Posts |
Posted - 2009-04-07 : 11:14:14
|
| [code]SELECT t.ComputerID,t.ClickDate,DATEDIFF(dd,t1.ClickDate,t.ClickDate) AS DiffFROM Table tOUTER APPLY(SELECT TOP 1 ClickDate FROM Table WHERE ComputerID=t.ComputerID AND ClickDate<t.ClickDate ORDER BY ClickDate DESC)t1[/code] |
 |
|
|
|
|
|
|
|